Dayton Real Estate Market EDA

Goals

• Look at relationships between features
• Establish a basic a understanding of the relationships
• Narrow down features to find ones that would be more useful in a deeper analysis
• Get ideas for a more involved analysis.

Load Packages

import gzip
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import warnings

import Redfin Metro Area Dataset. Contains Market Data since 2013

redfin_df = pd.read_csv(r"C:\Users\wilke\Downloads\redfin_metro_market_tracker.tsv000.gz",
                        compression='gzip',sep = '\t',header = 0,quotechar='"')

# get only Ohio records

oh_bool = []

for i in redfin_df['parent_metro_region']:
    bool = 'OH' in i
    oh_bool.append(bool)

oh_metrics = redfin_df[oh_bool]    

oh_metrics.sort_values(['parent_metro_region','period_begin'])

	period_begin	period_end	period_duration	region_type	region_type_id	table_id	is_seasonally_adjusted	region	city	state	...	sold_above_list_yoy	price_drops	price_drops_mom	price_drops_yoy	off_market_in_two_weeks	off_market_in_two_weeks_mom	off_market_in_two_weeks_yoy	parent_metro_region	parent_metro_region_metro_code	last_updated
13072	2012-01-01	2012-01-31	30	metro	-2	10420	f	Akron, OH metro area	NaN	NaN	...	-0.031343	NaN	NaN	NaN	0.032086	-0.017707	0.021275	Akron, OH	10420	2023-06-12 18:59:53
155928	2012-01-01	2012-01-31	30	metro	-2	10420	f	Akron, OH metro area	NaN	NaN	...	-0.125000	NaN	NaN	NaN	0.000000	0.000000	0.000000	Akron, OH	10420	2023-06-12 18:59:53
189842	2012-01-01	2012-01-31	30	metro	-2	10420	f	Akron, OH metro area	NaN	NaN	...	0.070000	NaN	NaN	NaN	0.000000	-0.071429	0.000000	Akron, OH	10420	2023-06-12 18:59:53
246828	2012-01-01	2012-01-31	30	metro	-2	10420	f	Akron, OH metro area	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Akron, OH	10420	2023-06-12 18:59:53
330225	2012-01-01	2012-01-31	30	metro	-2	10420	f	Akron, OH metro area	NaN	NaN	...	-0.035880	NaN	NaN	NaN	0.036364	-0.014095	0.024388	Akron, OH	10420	2023-06-12 18:59:53
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
213821	2023-04-01	2023-04-30	30	metro	-2	49780	f	Zanesville, OH metro area	NaN	NaN	...	-0.189678	0.223881	0.018752	0.118617	0.347222	0.019353	0.049925	Zanesville, OH	49780	2023-06-12 18:59:53
78549	2023-05-01	2023-05-31	30	metro	-2	49780	f	Zanesville, OH metro area	NaN	NaN	...	-0.086996	0.246753	0.039857	0.045570	0.400000	0.062162	0.034146	Zanesville, OH	49780	2023-06-12 18:59:53
154272	2023-05-01	2023-05-31	30	metro	-2	49780	f	Zanesville, OH metro area	NaN	NaN	...	0.000000	0.200000	NaN	NaN	0.666667	0.666667	0.666667	Zanesville, OH	49780	2023-06-12 18:59:53
312699	2023-05-01	2023-05-31	30	metro	-2	49780	f	Zanesville, OH metro area	NaN	NaN	...	-0.074074	0.244755	0.020875	0.032255	0.381579	0.034357	0.027149	Zanesville, OH	49780	2023-06-12 18:59:53
317645	2023-05-01	2023-05-31	30	metro	-2	49780	f	Zanesville, OH metro area	NaN	NaN	...	-1.000000	0.500000	NaN	NaN	1.000000	NaN	0.000000	Zanesville, OH	49780	2023-06-12 18:59:53

18561 rows × 58 columns

#All features in dataset

oh_metrics.columns

Index(['period_begin', 'period_end', 'period_duration', 'region_type',
       'region_type_id', 'table_id', 'is_seasonally_adjusted', 'region',
       'city', 'state', 'state_code', 'property_type', 'property_type_id',
       'median_sale_price', 'median_sale_price_mom', 'median_sale_price_yoy',
       'median_list_price', 'median_list_price_mom', 'median_list_price_yoy',
       'median_ppsf', 'median_ppsf_mom', 'median_ppsf_yoy', 'median_list_ppsf',
       'median_list_ppsf_mom', 'median_list_ppsf_yoy', 'homes_sold',
       'homes_sold_mom', 'homes_sold_yoy', 'pending_sales',
       'pending_sales_mom', 'pending_sales_yoy', 'new_listings',
       'new_listings_mom', 'new_listings_yoy', 'inventory', 'inventory_mom',
       'inventory_yoy', 'months_of_supply', 'months_of_supply_mom',
       'months_of_supply_yoy', 'median_dom', 'median_dom_mom',
       'median_dom_yoy', 'avg_sale_to_list', 'avg_sale_to_list_mom',
       'avg_sale_to_list_yoy', 'sold_above_list', 'sold_above_list_mom',
       'sold_above_list_yoy', 'price_drops', 'price_drops_mom',
       'price_drops_yoy', 'off_market_in_two_weeks',
       'off_market_in_two_weeks_mom', 'off_market_in_two_weeks_yoy',
       'parent_metro_region', 'parent_metro_region_metro_code',
       'last_updated'],
      dtype='object')

Heatmap to get a look at correlated variables

plt.figure(figsize = (20,20))
sns.heatmap(oh_metrics.corr())

C:\Users\wilke\AppData\Local\Temp\ipykernel_47604\1925643702.py:2: FutureWarning: The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.
  sns.heatmap(oh_metrics.corr())

<Axes: >

metrics to keep based on relationships

• median_list_price
• median_sale_price
• months_of_supply
• months_of_supply_mom
• months_of_supply_yoy
• homes_sold
• homes_sold_mom
• homes_sold_yoy
• median_list_ppsf
• median_list_ppsf_mom
• median_list_ppsf_yoy
• inventory
• inventory_mom
• inventory_yoy
• avg_sale_to_list
• sold_above_list
• sold_above_list_mom
• sold_above_list_yoy
• off_market_in_two_weeks
• price_drops
• pending_sales
• pending_sale_mom
• pending_sales_yoy

Histogram to visualize the distributions of all features

metrics_df = oh_metrics[[
    'period_begin',
    'parent_metro_region',
    'property_type',
    'median_list_price',
    'median_sale_price',
    'median_sale_price_mom',
    'median_sale_price_yoy',
    'months_of_supply',
    'months_of_supply_mom',
    'months_of_supply_yoy',
    'homes_sold',
    'homes_sold_mom',
    'homes_sold_yoy',
    'median_list_ppsf',
    'median_list_ppsf_mom',
    'median_list_ppsf_yoy',
    'inventory',
    'inventory_mom',
    'inventory_yoy',
    'avg_sale_to_list',
    'sold_above_list',
    'sold_above_list_mom',
    'sold_above_list_yoy',
    'off_market_in_two_weeks',
    'price_drops',
    'pending_sales',
    'pending_sales_mom',
    'pending_sales_yoy'
]]

#Visualize a histogram of all metrics to see distributions

ind = 1
plt.figure(figsize=(40,40))

for i in metrics_df.columns:
    plt.subplot(len(metrics_df.columns),2,ind)
    plt.hist(metrics_df[i])
    plt.title(i)
    ind += 1

• All features are heavily right skewed except price_drops and pending_sales_mom which is typical of real estate metrics

#Split dataset into dayton marekt

dayton_market = metrics_df[metrics_df['parent_metro_region'] == 'Dayton, OH']

Dayton Market EDA

dayton_townhouse = dayton_market[dayton_market['property_type'] == 'Townhouse'].sort_values('period_begin')
dayton_family_home = dayton_market[dayton_market['property_type'] == 'Single Family Residential'].sort_values('period_begin')
dayton_condo = dayton_market[dayton_market['property_type'] == 'Condo/Co-op'].sort_values('period_begin')
dayton_multi_family = dayton_market[dayton_market['property_type'] == 'Multi-Family (2-4 Unit)'].sort_values('period_begin')
dayton_residential = dayton_market[dayton_market['property_type'] == 'All Residential'].sort_values('period_begin')

Time-Series look at metrics of each specific property type

Townhouses

metrics = [ 'median_sale_price',
            'inventory',
            'median_list_price',
            'homes_sold',
            'avg_sale_to_list',
            'sold_above_list',
            'homes_sold',
            'median_list_ppsf',
            'pending_sales',
            'off_market_in_two_weeks',
            'price_drops',
            ]



fig, ax = plt.subplots(11,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics):
    ax[ind].plot(dayton_townhouse['period_begin'],dayton_townhouse[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_townhouse['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_42968\3672131972.py:25: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_townhouse['period_begin'],rotation = 90)

Single Family Homes

metrics = [ 'median_sale_price',
            'inventory',
            'median_list_price',
            'homes_sold',
            'avg_sale_to_list',
            'sold_above_list',
            'homes_sold',
            'median_list_ppsf',
            'pending_sales',
            'off_market_in_two_weeks',
            'price_drops',
            ]



fig, ax = plt.subplots(11,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics):
    ax[ind].plot(dayton_family_home['period_begin'],dayton_family_home[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_family_home['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_42968\4127881422.py:31: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_family_home['period_begin'],rotation = 90)

Condo/Co-op

metrics = [ 'median_sale_price',
            'inventory',
            'median_list_price',
            'homes_sold',
            'avg_sale_to_list',
            'sold_above_list',
            'homes_sold',
            'median_list_ppsf',
            'pending_sales',
            'off_market_in_two_weeks',
            'price_drops',
            ]



fig, ax = plt.subplots(11,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics):
    ax[ind].plot(dayton_condo['period_begin'],dayton_condo[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_condo['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_42968\1762790433.py:25: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_condo['period_begin'],rotation = 90)

Muti-Family (2-4 Units)

metrics = [ 'median_sale_price',
            'inventory',
            'median_list_price',
            'homes_sold',
            'avg_sale_to_list',
            'sold_above_list',
            'homes_sold',
            'median_list_ppsf',
            'pending_sales',
            'off_market_in_two_weeks',
            'price_drops',
            ]



fig, ax = plt.subplots(11,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics):
    ax[ind].plot(dayton_multi_family['period_begin'],dayton_multi_family[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_multi_family['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_42968\2919694956.py:25: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_multi_family['period_begin'],rotation = 90)

All Residential

metrics = [ 'median_sale_price',
            'inventory',
            'median_list_price',
            'homes_sold',
            'avg_sale_to_list',
            'sold_above_list',
            'homes_sold',
            'median_list_ppsf',
            'pending_sales',
            'off_market_in_two_weeks',
            'price_drops',
            ]



fig, ax = plt.subplots(11,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics):
    ax[ind].plot(dayton_residential['period_begin'],dayton_residential[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_residential['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_42968\3061486996.py:25: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_residential['period_begin'],rotation = 90)

Takeaways from time-series graphs

• Dayton being a primarily residential/single-family area provides more consistent broad market trends.
• Condos, multifamily, and townhomes being less common, the market trends are less predictable from a broad market standpoint as they have a smaller sample.
• There has been a fairly large increase in homes being sold above list signifying a competitve buyers market.

Month-Over-Month Metrics

Only going to look at residential and single-family as they follow more predictable trends

Single-Family Home

Month_over_month

metrics_mom = [ 'median_sale_price_mom',
            'inventory_mom',
            'homes_sold_mom',
            'sold_above_list_mom',
            'median_list_ppsf_mom',
            'pending_sales_mom'
            ]



fig, ax = plt.subplots(6,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics_mom):
    ax[ind].plot(dayton_family_home['period_begin'],dayton_condo[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_family_home['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_47604\759831397.py:20: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_family_home['period_begin'],rotation = 90)

All Residential

Month-Over-Month

metrics_mom = [ 'median_sale_price_mom',
            'inventory_mom',
            'homes_sold_mom',
            'sold_above_list_mom',
            'median_list_ppsf_mom',
            'pending_sales_mom'
            ]



fig, ax = plt.subplots(6,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics_mom):
    ax[ind].plot(dayton_residential['period_begin'],dayton_residential[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_residential['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_47604\1691309848.py:20: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_residential['period_begin'],rotation = 90)

Month-Over-Month Observations

• Since the real estate market is seasonal, month over month doesn't provide as much insight as present metrics and year-over-year metrics.

Year-Over-Year Metrics

Single-Family Home

metrics_mom = [ 'median_sale_price_yoy',
            'inventory_yoy',
            'homes_sold_yoy',
            'sold_above_list_yoy',
            'median_list_ppsf_yoy',
            'pending_sales_yoy'
            ]



fig, ax = plt.subplots(6,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics_mom):
    ax[ind].plot(dayton_family_home['period_begin'],dayton_condo[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_family_home['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_47604\2833583520.py:20: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_family_home['period_begin'],rotation = 90)

All Residential

Year-Over-Year

metrics_mom = [ 'median_sale_price_yoy',
            'inventory_yoy',
            'homes_sold_yoy',
            'sold_above_list_yoy',
            'median_list_ppsf_yoy',
            'pending_sales_yoy'
            ]



fig, ax = plt.subplots(6,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics_mom):
    ax[ind].plot(dayton_residential['period_begin'],dayton_residential[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_residential['period_begin'],rotation = 90)

C:\Users\wilke\AppData\Local\Temp\ipykernel_47604\1910253905.py:20: UserWarning: FixedFormatter should only be used together with FixedLocator
  ax[ind].set_xticklabels(dayton_residential['period_begin'],rotation = 90)

Year-Over-Year Observations

• Sold_above_list and home_sold have seen year over year declines. Inventory has recent spikes inventory year-over-year growth.
  • indicating decreasing market competition

Implement Historical Mortgage Rate Data

Import dataset from FreddieMac

mortgage_df = pd.read_csv('mortgageData.csv',encoding= 'unicode_escape')

#convert to datetime

mortgage_df['Week'] = pd.to_datetime(mortgage_df['Week'])

#get only records that match with redfindata

mortgage_df = mortgage_df[mortgage_df['Week'] >= '2012-01-01']

#create columns to get dates groupable by month

warnings.filterwarnings("ignore")

mortgage_df['year'] = mortgage_df['Week'].dt.year
mortgage_df['month'] = mortgage_df['Week'].dt.month
mortgage_df['day'] = [1 for i in range(len(mortgage_df))]

mortgage_df

	Week	30YrFRM	30YrFees	15YrFRM	15YrFees	51ARM	51ARMFees	year	month	day
2127	2012-01-05	3.91	0.8	3.23	0.8	2.86	0.7	2012	1	1
2128	2012-01-12	3.89	0.7	3.16	0.8	2.82	0.7	2012	1	1
2129	2012-01-19	3.88	0.8	3.17	0.8	2.82	0.7	2012	1	1
2130	2012-01-26	3.98	0.7	3.24	0.8	2.85	0.7	2012	1	1
2131	2012-02-02	3.87	0.8	3.14	0.8	2.80	0.7	2012	2	1
...	...	...	...	...	...	...	...	...	...	...
2721	2023-05-25	6.57	NaN	5.97	NaN	NaN	NaN	2023	5	1
2722	2023-06-01	6.79	NaN	6.18	NaN	NaN	NaN	2023	6	1
2723	2023-06-08	6.71	NaN	6.07	NaN	NaN	NaN	2023	6	1
2724	2023-06-15	6.69	NaN	6.10	NaN	NaN	NaN	2023	6	1
2725	2023-06-22	6.67	NaN	6.03	NaN	NaN	NaN	2023	6	1

599 rows × 10 columns

#Create a new date column with days of the month all the same
mortgage_df['Date'] = pd.to_datetime({'year':mortgage_df['year'],
                                      'month':mortgage_df['month'],
                                      'day':mortgage_df['day']})

#drop original date columns
mortgage_clean_dates = mortgage_df.drop(['Week','year','month','day'],axis = 1)

#group by month in order to be able to join on redfin data

thirtyFRM = pd.DataFrame(mortgage_clean_dates.groupby('Date').agg(np.mean)['30YrFRM']).reset_index()
thirtyFRM = thirtyFRM[:-1]

dayton_residential.reset_index(inplace = True)

Join Redfin market data and Freddie Max mortgage data

dayton_residential_FRM = dayton_residential.merge(right = thirtyFRM['30YrFRM'], how = 'inner',right_index=True,left_index=True)
dayton_residential_FRM.columns

Index(['index', 'period_begin', 'parent_metro_region', 'property_type',
       'median_list_price', 'median_sale_price', 'median_sale_price_mom',
       'median_sale_price_yoy', 'months_of_supply', 'months_of_supply_mom',
       'months_of_supply_yoy', 'homes_sold', 'homes_sold_mom',
       'homes_sold_yoy', 'median_list_ppsf', 'median_list_ppsf_mom',
       'median_list_ppsf_yoy', 'inventory', 'inventory_mom', 'inventory_yoy',
       'avg_sale_to_list', 'sold_above_list', 'sold_above_list_mom',
       'sold_above_list_yoy', 'off_market_in_two_weeks', 'price_drops',
       'pending_sales', 'pending_sales_mom', 'pending_sales_yoy', '30YrFRM'],
      dtype='object')

dayton_residential_FRM

	index	period_begin	property_type	parent_metro_region	median_sale_price	median_sale_price_mom	median_sale_price_yoy	median_list_price	median_list_price_yoy	homes_sold	...	avg_sale_to_list	sold_above_list	sold_above_list_mom	sold_above_list_yoy	off_market_in_two_weeks	price_drops	pending_sales	pending_sales_mom	pending_sales_yoy	30YrFRM
0	155893	2012-01-01	All Residential	Dayton, OH	69900.0	-0.115190	-0.068000	80000.0	-0.089875	455.0	...	0.918299	0.105495	-0.002525	0.012944	0.260450	NaN	311.0	0.187023	0.126812	3.9150
1	132602	2012-02-01	All Residential	Dayton, OH	75000.0	0.072961	-0.025974	92000.0	-0.023355	533.0	...	0.934589	0.133208	0.027714	0.031759	0.305040	NaN	377.0	0.212219	0.455598	3.8900
2	120193	2012-03-01	All Residential	Dayton, OH	85000.0	0.133333	0.223022	110000.0	0.004566	725.0	...	0.935872	0.097931	-0.035277	-0.043789	0.393814	NaN	485.0	0.286472	0.203474	3.9540
3	237560	2012-04-01	All Residential	Dayton, OH	97000.0	0.141176	0.197531	99900.0	-0.074143	702.0	...	0.942678	0.119658	0.021727	0.029028	0.329710	NaN	552.0	0.138144	0.157233	3.9100
4	336108	2012-05-01	All Residential	Dayton, OH	95570.0	-0.014742	0.061889	107900.0	0.079011	840.0	...	0.949053	0.120238	0.000580	-0.014258	0.270796	NaN	565.0	0.023551	0.304850	3.7980
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
132	331695	2023-01-01	All Residential	Dayton, OH	180000.0	0.000000	0.094225	169900.0	0.000000	515.0	...	0.970267	0.207767	-0.020370	-0.114142	0.429231	0.381927	650.0	0.116838	-0.203431	6.2725
133	255251	2023-02-01	All Residential	Dayton, OH	180000.0	0.000000	0.033592	185500.0	0.031128	673.0	...	0.973408	0.233284	0.025517	-0.160159	0.503279	0.254593	610.0	-0.061538	-0.188830	6.2575
134	301717	2023-03-01	All Residential	Dayton, OH	185000.0	0.027778	0.054131	199900.0	0.000000	811.0	...	0.983608	0.311961	0.078677	-0.145648	0.571770	0.254919	836.0	0.370492	-0.130073	6.5440
135	180461	2023-04-01	All Residential	Dayton, OH	200000.0	0.081081	0.000000	209900.0	0.105319	838.0	...	0.993206	0.403341	0.091381	-0.118170	0.558580	0.241220	845.0	0.010766	-0.211018	6.3425
136	113081	2023-05-01	All Residential	Dayton, OH	212735.0	0.063675	0.037732	219900.0	0.072683	987.0	...	0.998522	0.436677	0.033336	-0.079396	0.540948	0.268335	928.0	0.098225	-0.247974	6.4250

137 rows × 31 columns

A look at home sale price and inventory, and their relationship to mortgage rates

sales_FRM = dayton_residential_FRM[['30YrFRM','median_sale_price']]
inventory_FRM = dayton_residential_FRM[['30YrFRM','inventory']]

plt.scatter(sales_FRM['median_sale_price'],sales_FRM['30YrFRM'])
plt.xlabel('Median Sale Price')
plt.ylabel('Mortgage Rate')

print(f'Correlation Coefficient: {sales_FRM.corr()}')

Correlation Coefficient:                     30YrFRM  median_sale_price
30YrFRM            1.000000           0.262442
median_sale_price  0.262442           1.000000

plt.scatter(inventory_FRM['inventory'],inventory_FRM['30YrFRM'])
plt.xlabel('Inventory')
plt.ylabel('Mortgage Rate')

print(f'Correlation Coefficient: {sales_FRM.corr()}')

Correlation Coefficient:                     30YrFRM  median_sale_price
30YrFRM            1.000000           0.262442
median_sale_price  0.262442           1.000000

metrics_mom = [ 'median_sale_price',
            'inventory',
            '30YrFRM'
            ]



fig, ax = plt.subplots(3,1,sharex = True,figsize = (20,30))

fig.tight_layout(pad = 5)

ax = ax.flatten()

for ind,i in enumerate(metrics_mom):
    ax[ind].plot(dayton_residential_FRM['period_begin'],dayton_residential_FRM[i])
    ax[ind].set_title(i)
    ax[ind].set_xticklabels(dayton_residential_FRM['period_begin'],rotation = 90)

Key Takeaways

- There is a an inverse relationship between inventory and price. Basic Supply and Demand Structure.

Questions about this

• When does inventory/supply bottom out?
  • When mortgage rates are low, incentive to add to inventory via developers.
  • When mortgage rate are high, developers are less incentivized to build.
  • In final graph, inventories seasonal winter trench was higher this year than year before.
• What drives inventory?
  • Analysis doesn't show a strong relationship between inventory and mortgage rates.
  • Outside of recent years, mortgage rates have been relatively stable.
  • Intuition is that higher mortgage rates correlate to higher inventory (less people are wanting to buy when its more expensive)
  • Inventory could have a very low elasticity to mortgage rates (no relationship seen when mortage rates were stable) but will this spike in mortgage rates finally show an effect on inventory and how long would the lag be between mortgage rate spike and increase in inventory.

- Townhome market is more unstable (most likely due to smaller sample)

- Scatterplots show these mortgage rates are clear anomalies in past 10 years in relation to where pricing and inventory are at

Next Steps

• Incorporate Businesses and Restaurants being built in area and look at relationships there.
• Forecasting Price, Inventory using the auto-regression with exogenous variables found in analysis.